blender return wrong value for matrix #59726

New Issue

Francisco Javier Martinez Palmer · 2018-12-21T16:47:15+01:00

Francisco Javier Martinez Palmer commented

2018-12-21 16:47:15 +01:00

System Information
Operating system: windows
Graphics card: nvidia quadro

Blender Version
Broken: 2.79b release

Short description of error
blender is providing wrong values for the pose matrix when getting them.
The top bone is rotated 60 degrees on x and y axis, and moved up around 20 units.
The matrix for that bone is correct as its displayed in the matrix (check the python code provided in the scene)
When trying to retrieve the values from the matrix, some values are wrong. In this case one specifically (check the code as mentioned to know more)

Exact steps for others to reproduce the error
checkout the provided scene.
Steps to reproduce:

Select the box mesh (is very distorted)
Run the script Text.001

Now check the script code and the output in the console. No mather if deepcopy or anything else is used blender always returns different values than the ones stored in the matrix.repo_cube.blend

**System Information** Operating system: windows Graphics card: nvidia quadro **Blender Version** Broken: 2.79b release **Short description of error** blender is providing wrong values for the pose matrix when getting them. The top bone is rotated 60 degrees on x and y axis, and moved up around 20 units. The matrix for that bone is correct as its displayed in the matrix (check the python code provided in the scene) When trying to retrieve the values from the matrix, some values are wrong. In this case one specifically (check the code as mentioned to know more) **Exact steps for others to reproduce the error** checkout the provided scene. Steps to reproduce: - Select the box mesh (is very distorted) - Run the script Text.001 Now check the script code and the output in the console. No mather if deepcopy or anything else is used blender always returns different values than the ones stored in the matrix.[repo_cube.blend](https://archive.blender.org/developer/F6026620/repo_cube.blend)

Francisco Javier Martinez Palmer commented

2018-12-21 16:47:15 +01:00

Added subscriber: @matheboy

Sebastian Parborg commented

2018-12-21 17:52:25 +01:00

Added subscriber: @ZedDB

Sebastian Parborg commented

2018-12-21 17:52:25 +01:00

Changed status from 'Open' to: 'Archived'

Sebastian Parborg closed this issue

2018-12-21 17:52:25 +01:00

Sebastian Parborg self-assigned this 2018-12-21 17:52:25 +01:00

Sebastian Parborg commented

2018-12-21 17:52:25 +01:00

The values are correct as the matrix print does not print out the whole float value. The later prints outputs the actual value.

IE if the matrix value is less than 0.0001, it will print it as 0.0000. However, that doesn't mean that the actual floating point value is 0

The values are correct as the matrix print does not print out the whole float value. The later prints outputs the actual value. IE if the matrix value is less than 0.0001, it will print it as 0.0000. However, that doesn't mean that the actual floating point value is 0

Francisco Javier Martinez Palmer commented

2018-12-23 20:43:16 +01:00

Changed status from 'Archived' to: 'Open'

Francisco Javier Martinez Palmer reopened this issue

2018-12-23 20:43:16 +01:00

Francisco Javier Martinez Palmer commented

2018-12-23 20:43:16 +01:00

This are the lines you should be looking at (starting at 36 in the script):

      print (aux[0]) # here the vector is : <Vector (0.5000, -0.0000, -0.8660, 0.0000)>
      aux3 = copy.deepcopy(aux[0])
      print(aux3)
      print(aux3[0],aux3[1])
      print (aux[0].y) # why here then returns -1.3461173509199398e-08 ???? it should be -0.0000 !!!!!!

Do you think the difference between 0.0 (in the matrix) and -1.34 (when getting the value out) is acceptable?

If you look closer this happens later again with the other matrix, it doesnt look acceptable to me

This are the lines you should be looking at (starting at 36 in the script): ``` print (aux[0]) # here the vector is : <Vector (0.5000, -0.0000, -0.8660, 0.0000)> aux3 = copy.deepcopy(aux[0]) print(aux3) print(aux3[0],aux3[1]) print (aux[0].y) # why here then returns -1.3461173509199398e-08 ???? it should be -0.0000 !!!!!! ``` Do you think the difference between 0.0 (in the matrix) and -1.34 (when getting the value out) is acceptable? If you look closer this happens later again with the other matrix, it doesnt look acceptable to me

Stephen Swaney commented

2018-12-23 20:55:06 +01:00

Added subscriber: @StephenSwaney

Stephen Swaney commented

2018-12-23 20:55:06 +01:00

Do you think the difference between 0.0 (in the matrix) and -1.34 (when getting the value out) is acceptable?

If I may...
The return value is not -1.34, but -1.3461173509199398e-08.
Printed in non-scientific notation it looks like -0.00000001.346 (rounded).
Mathematically speaking, this is DNZ - Damn Near Zero.

Internally, calculations are done with the full floating point values, when printed they get rounded

> Do you think the difference between 0.0 (in the matrix) and -1.34 (when getting the value out) is acceptable? If I may... The return value is not -1.34, but -1.3461173509199398e-08. Printed in non-scientific notation it looks like -0.00000001.346 (rounded). Mathematically speaking, this is DNZ - Damn Near Zero. Internally, calculations are done with the full floating point values, when printed they get rounded

Sebastian Parborg commented

2018-12-23 21:44:36 +01:00

Changed status from 'Open' to: 'Archived'

Sebastian Parborg closed this issue

2018-12-23 21:44:36 +01:00

Sebastian Parborg commented

2018-12-23 21:44:36 +01:00

@StephenSwaney Thanks for the explaination

@matheboy In the future, do not reopen issues just because you do not understand the answer. You are of course free to ask for a clarification. I incorrectly assumed that you knew the standard scientific notations and thought you were confused about floating point numbers. (IE why is it near zero when it should be zero?)

https://en.wikipedia.org/wiki/Scientific_notation#E-notation

@StephenSwaney Thanks for the explaination @matheboy In the future, do not reopen issues just because you do not understand the answer. You are of course free to ask for a clarification. I incorrectly assumed that you knew the standard scientific notations and thought you were confused about floating point numbers. (IE why is it near zero when it should be zero?) https://en.wikipedia.org/wiki/Scientific_notation#E-notation

Sign in to join this conversation.

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

blender return wrong value for matrix #59726